Reactions of transition-metal atoms with arenes, arene-functionalized

Wrapping oligomers and polymers around metal atoms, metal clusters, and metal colloids. Mark P. Andrews and Geoffrey A. Ozin. Chemistry of Materials 1...
16 downloads 0 Views 1MB Size
Inorg. Chem. 1986, 25, 2587-2595 those of the present study. The retardation effect by addition of free base L can be interpreted by the decrease in concentration of U02(acac),acacH with the addition of L. This interpretation is similar to that of the retardation in acac exchange in Th(acac), but different from that in the exchange of acac in VO(acac),,'2 where the retardation is attributed to the formation of an outer-sphere intermediate with the added base. As mentioned above, the present substitution reaction in UOz(acac)zL and the exchange reaction of acac in Th(acac), proceed through the type 3 mechanism, where the ratedetermining step is the proton-transfer process. Although there is a similarity between the mechanisms of both systems, the reaction rates differ considerably from each other. The difference in rates of the proton-transfer processes can be understood by assuming that the process consists of three steps:

where Mn+ is a metal ion and HO-0 represents a 8-diketonate (HB) in the enol form. In the first step in eq 7, incoming HB* and coordinated B- come near each other to make a hydrogen bridge between HB* and B-. The second step is the proton transfer

2587

including the formation of the hydrogen bridge. The third step is the breaking of the hydrogen bridge between coordinated B-* and leaving HB. The difference in rates is mainly attributed to the difference in AG*values for the first step, since the variation in AG*values for the proton-transfer step is expected to be relatively small. The unidentate P-diketonate as a proton donor, which was coordinated to the metal ion with a large ionic radius, rotates about a metal-oxygen bond so freely that the 8-diketonate can easily orientate so as to favor proton transfer. Large negative AS*values (Tables I1 and 111) seem to reflect the steric hindrance in the first step. Consequently, the difference in rates between U 0 , ( a ~ a c ) ~and L Th(acac), or U(acac),' is explained by the difference in ionic radii of central metal ions. The steric hindrance may be responsible fin part for the difference in rates of exchange between Th(acac), and Th(tta),, (tta = 2-thionyltrifluoroacetonate) because tta is more bulky than acac. Although many studies on the intermolecular proton-transfer reaction have been repported, data on the rate of the intramolecular proton-transfer reaction, particularly in organic solvents, are not available. More detailed studies on the U02(P-diketonate)2L system are in progress in order to elucidate the steric hindrance and kinetic isotope effect in the intramolecular proton-transfer process, and information on these aspects is to be considered in a future report. Acknowledgment. W.-S.J. wishes to thank Dr. Y. Ikeda for providing uranyl complexes. Registry No. U02(acac)2(Me2SO),71357-22-7; U02(acac)*(dmf), 89145-1 1-9; U02(acac),(tmp), 102575-53-1; dbm, 19274-26-1; D2, 7782-39-0.

Contribution from the Lash Miller Chemical Laboratories, University of Toronto, Toronto, Ontario, Canada M5S 1Al

Reactions of Transition-Metal Atoms with Arenes, Arene-Functionalized Alkanes, Oligo(ethylene oxides), and Polysiloxanes. 4 Mark P. Andrewst and Geoffrey A. Ozin* Received October 18, 1985 New synthetic, chemical, and spectroscopic (optical, Raman, IR, EPR) experiments aimed at further elucidation of the details of the near-room-temperature reactions of metal atoms with arene-functionalizedliquid polysiloxanes to produce organometallic polymers and polymer-supported metal clusters are described. By examining the reactions of V, Cr,and Mo atoms with model arenes, arene-functionalized alkanes, and arene-functionalizedoligo(ethy1ene oxides), one is able to gain insight into the role of arene substituents, bis(arene)metal cross-links, and the solvating backbone in the mode of formation and stabilization of metal cluster sites on the polysiloxane. The chemical and physical homogeneity of the resulting polymers is also assessed.

Introduction A theme running through much of our recent metal atom research has been to find ways of generating and stabilizing metal clusters at or close to room temperature in various fluid trapping media and to devise methods for investigating their physical and chemical properties within the trapping medium itself.' The ability to transfer the metal aggregates from the fluid intact to supports such as carbon, inorganic oxides, and zeolites emerged as a result of such investigations.**2 In this paper we focus attention on a method of metal cluster formation in the liquid phase that is based on the direct sequential addition of metal atoms according to sequence 1 during the metal M + Mn

+

%+I

(1)

atom deposition process. This reaction is viewed as occurring at a site preformed in the support medium, which invariably contains arene functionalities, usually linked to oligoalkanes, oligo(ethy1ene oxides), and oligo- and polysiloxanes. In particular, the arene substituents inhibit competing metal atom polymerization reactions Current address: AT&T Bell Laboratories, Murray Hill, NJ 07974.

0020-1669/86/1325-2587$01.50/0

Scheme I

\-

MtSi

by chemically trapping the atomic diffusant. In general, these liquid supports permit chemical reactions of metal atoms with ligands to occur at higher temperatures, which will result in a more favorable competitive balance between bis(arene)M complex (1) Ozin, G. A. CHEMTECH 1985, I S , 488 and references cited therein. (2) Ozin, G. A.; Andrews, M. P. US.Patent 454886, 1982. Ozin, G. A.; Hugues, F.;Nazar, L. F. US. Patent 454886, 1982. Nazar, L. F.; Ozin, G. A,; Hugues, F.; Godber, J.; Rancourt, D. J . Mol. Catal. 1983, 21, 313. Ozin, G. A.; Andrews, M. P. Stud. Surf.Sei. Catal., in press.

0 1986 American Chemical Society

2588 Inorganic Chemistry, Vol. 25, No. 15, 1986

Andrews and Ozin

W

ip

\.

2

L

4

8

6

.

IO

. I2

X I

Figure 2. Graphical representation of a Mo/DC 510 reaction at 250 K of the type shown in ref 4.

-

I

I

250

300

I

1

I

350

400

450

Figure 1. UV-vis spectra: A, (C6HSCH3)2M~ in spectral grade pentane, 3.4 X 10" M (e = 1.1 X lo5,X = 318 nm); B, (C6H5CH3)2M~ in poly(dimethylsi1oxane) (DC 200) fluid; C, the (arene)2Mo complex prepared in situ in poly(dimethy1-co-methylphenylsiloxane)(DC 5 10) by electron-beam or resistive metal vaporization.

formation and.meta1 atom agglomeration, other things being equal. Francis and Timms3 first reported that bis(arene) transitionmetal complexes could be formed by cocondensing transition-metal atoms of Ti, V, Cr, Mo, and W into liquid poly(methylpheny1siloxanes, DC 510) at 273 K using metal vapor rotary reaction techniques. These reactions produce organometallic polymers as shown in Scheme I. The principle of these reactions has been extended to quantitative depositions of transition-metal atoms into thin, quiescent liquid films of these nonvolatile materials, particularly the Dow Coming 5 10 fluid."8 The thin-film experiments form the intermediate link between the matrix-isolation experiments and the macroscale rotary reactor experiments. The present study is organized as follows: We very briefly review published results of liquid-thin-film experiments involving reactions of transition-metal atoms with the poly(methy1phenyl-co-dimethylsiloxane) (DC 5 10) system. The experiments and arguments implicating metal cluster compounds in this macromolecular support have been detailed in earlier papers from this laboratory to which the reader is referred."* In this paper, some further insight into the mode of stabilization of metal clusters in the DC 510 system is obtained from studies involving different phenyl-substituted monomers and oligomers in metal vapor reactions surveying the temperature range 12-300 K. Experimental Section The experimental details of metal vaporization into siloxane polymers containing pendant arena have been described previously.'" Quantita-

(3) Francis, C.G.; Timms, P. L. J . Chem. Soc., Dalton Trans. 1980,1401. (4) Ozin, G. A.; Francis, C. G. J. Mol. Srrucr. 1 9 0 , 9, 55. (5) Francis, C.G.; Ozin, G. A. J. Macromol. Sci., Chem. A16, 1981,167. (6) Francis, C.G.; Huber, H.; Ozin, G. A. J. Am. Chem. SOC.1979,101, 6250. (7) Francis, C.G.; Huber, H.; Ozin, G. A. Inorg. Chem. 1980, 19, 219. (8) Francis, C.G.; Huber, H.; Ozin, G. A. Angew. Chem., Int. Ed. Engl. 1980, 19, 402.

0.4

0

2 0.3

P

0.i

0.1

C

/ I

i

I

i

I

i

I

i

6

i2

x

Figure 3. Plot of AMcon/AMoshowing independent growth characteristics. tive studies of metal atom reactions in these polymers were made possible by the development of a static, liquid-thin-film microscale analogue to the macrosynthetic experiments? In brief, the method exploits the characterization tools developed for matrix-isolation experiments and was thus suitably adapted for in situ microgram-scale metal vapor synthesis and spectroscopy.@ In what follows we detail only those aspects of the chemical and spectroscopic procedures that are unique to the present investigation. Synthetic Procedures. (a) Arene End-Functionalized Oligo(ethy1ene oxides). Preparations of these compounds were based on the methods of Perry and Hibbert'O and on the method of Bomer et al." For experiments symbolized by Figure 4, the a,w-diphenyl ethers of tetrathrough heptaethylene glycol were synthesized. Tetraethylene glycol was purchased from Aldrich. The remaining glycols were synthesized according to the method prescribed for hexaethylene glycol. (i) Hexaethylene Glycol. The a,w-dibromo derivative of diethylene glycol was synthesized according to the procedure of Dann et a1.I2in 40% yield. The monosodium salt of diethylene glycol (dried over activated

(9) Ozin, G. A.;Francis, C. G.; Huber, H. X.;Andrews, M.; Nazar, L. F. J . Am. Chem. Soc. 1981,103,2453. Hibbert, H. Can. J. Res., Sect. B 1936, 14, 77. (10) Perry, S.Z.; (11) Bomer, B.; Heitz, W.; Kern, J. J . Chromatogr. 1970,53, 51. (12) Dann, J. R.;Chisea, P. P.; Gates, J., Jr. J. Chem. SOC.1961, 1991.

Reactions of Transition-Metal Atoms with Polysiloxanes

M -

f L I I I

I

‘,

‘\- -&-%W!YE_S- - - - - - - - - - +vl

Figure 4. Matrix and macroscale metal vapor synthetic strategy used to elucidate the nature of metal cluster diffusion/agglomeration/complexation in DC 510. 4A molecular sieves) was prepared as prescribed by Bomerl’ with freshly cut sodium metal, all manipulations being made under a nitrogen atmosphere until the alkoxide formation was complete (several hours). Thus to 600 g (5.66 mol) of mechanically stirred diethylene glycol was added 184.2 g (1.88 mol) of sodium metal. The solution turned orange and became quite viscous during the room-temperature reaction. a,wDibromodiethyl ether (227.7 g, 0.99 mol) was added slowly to the stirred alkoxide solution. NaBr precipitated, and the solution was heated to 75 OC until it became neutral to pH paper (ca. 20 min). The orange-brown solution was separated from the salt by vacuum filtration through a coarse sintered-glass funnel. This liquid was still contaminated with NaBr and was finally freed of the salt by liquid-liquid extraction (Et20) after 5 days. A pale yellow extract collected in the ether pot. The ether was removed under vacuum, and the viscous bright yellow oil distilled through a 15-cm Vigreux column (101.2 g, 178 OC (0.05 mm), lit.lo bp 203-205O (0.03 mm)). The hexamethyldisilazane derivative of the diol was prepared and chromatographed according to the method of Calzolari et al.,l3 which indicated that the diol was pure. NMR (CDCI,, 60 MHz): 6 3.01 (s, 2 H), 3.65 (s, 24 H). IR (neat): 3540 cm-I (br, OH bonded). (ii) Hexaethylene Glycol Ditosylate. Hexaethylene glycol (8.5 g, 0.03 mol) was dissolved in 25-mL dried (CaO) pyridine and cooled to -5 OC. Tosyl chloride (1 1.8 g, 0.062 mol) in 20 mL of pyridine was added with stirring at such a rate that the temperature of the reaction mixture did not rise above -2 “C (formation of pyridinium chloride precipitate). The reaction was stirred for 18’/2 h. The reaction mixture was poured over 300 mL of crushed ice/water and extracted with CH2CI2(3 X 100 mL). The combined organic phases were washed with 3 X 300 mL of ice cold 6N HCI and 250 mL of cold, saturated ammonium chloride. The organic phase was dried over anhydrous MgS04 and filtered, and the solvent was removed under vacuum to yield a faintly yellow-brown colored oil after chromatography (silica, EtOH:CH2CI2 2%:98% v/v). Yield: 16.53 g, 93% NMR (CDCI,, 60 MHz): 6 2.40 (br s, 6 H, ArCH,), 3.57 (s, 8 H, -OCH2CH20-), 3.61 (s, 8 H, -OCH2CH20-), 3.5-3.8 (m, 4 H, TsOCH2CH20-), 3.97-4.28 (m, 4 H, TsOCH2CH20-), 7.29 and 7.72 (AA’BB’, 8 H, J A B = 8 Hz, ArH). (i) o,wDiphenyl Ether of Hexaethylene Glycol. The ditosylate (1 1.84 g) in 10 mL of T H F was added to 3.76 g (0.04 mol) of phenolate anion prepared by adding 2.3 g (0.041 mol) of KOH dissolved in 2.5 mL of H 2 0 to the phenol in 65 mL of THF. The solution was stirred under N2 at 70 OC for 5 h and then at 45 OC for 16 h. When the liquid was cooled, in ice, a large mass of white crystals deposited. These were separated by vacuum filtration. The T H F was removed under vacuum and the (13) Calzolari, C.; Stancher, B.; Favretto, L. J . Chromatogr. 1968, 38, 7. (14) Tummler, B.;Maass, G.; Vogtle, F.; Sieger, H.; Heimann, U.; Weber, E. J. Am. Chem. SOC.1979, 101, 2588.

Inorganic Chemistry, Vol. 25, No. 15, 1986 2589 yellow oil chromatographed on alumina (Et20:pentane, 2:3). The product was pale yellow and could not be decolorized. NMR (CDCI,, 60 MHz): b 3.58 (s, 8 H, -OCH*CH20-), 3.63 (s, 8 H, -OCH2CH20-), 3.5-3.9 (m, 4 H, ArOCH2CH20-), 3.96-4.16 (m, 4 H, ArOCH2CH20-), 6.64-7.37 (m, 10 H, ArH). The a,w-diphenyl ethers of the tetra-, penta-, and heptaethylene glycols were similarly prepared. (iv) a,w-DiphenyltetraethyleneGlycol. NMR (acetone-d,, 60 MHz): b 3.61 (s, 4 H, -OCH,CH20-), 3.63 (s, 4 H, -OCH2CH20-), 3.60-3.90 (m, 4 H, ArOCH2CH20-), 4.0-4.23 (m, 4 H, ArOCH2CH20-), 6.70-7.42 (m, 10 H, ArH). Lit. NMR (CDCI,): 6 3.68-4.2 (m, 16 H, -OCH2CH20-). (v) a,o-DiphenylpentaethyleneGlycol. NMR (acetone-d,, 60 MHz): 6 3.58-4.23 (m, 20 H, -OCH2CH20-), 6.78-7.42 (m. 10 H, ArH). (vi) a,w-DitoluylheptaethyleneGlycol. NMR (acetone-d,, 60 MHz): 6 2.23 (s, 6 H, ArCH,), 3.50-4.21 (m, 28 H, -OCH2CH20-), 6.68-7.20 (m, 8 H, ArH). (vii) Microscale Metal Vapor Reactions of the a,w-Diaryloligo(ethylene glycols). Reactions of these liquids were carried out at or near room temperature according to the method described in ref 4. (a) Synthesis of a,o-Ditoluyldecane. This compound was prepared by the Fittig reaction between p-bromotoluene and 1,lO-dibromodecane (Caution! 1,lO-dibromodecane is a suspected potent carcinogen). The product was purified by preparative GLC (support, Chromosorb W HP, 80/100 mesh, OV-101 coated, 6 ft X 2 mm, 200 “C). NMR (acetone-d6, 60 MHz): b 1.2-2.0 (br m, 12 H, -CH2), 2.4-2.82 (m, 8 H, Ar(CH2CH2-)), 2.3 (s, 6 H, ArCH,), 7.13 (s, 8 H, ArH). (i) Microscale Metal Vapor Reactions of a,w-Ditoluyldecane. This compound is a solid at room temperature. The following method was devised to carry out a solution-phase metal atom reaction: A 15% w/v solution of the ligand in Fischer (125/135, arene free) paraffin oil was mixed with 1 mL of spectral grade hexane and applied as a thin film, via vacuum microsyringe, to a horizontal NaCl window as described in ref 9. The hexane quickly evaporated, leaving behind an unbroken film of liquid. Vanadium or molybdenum atoms were allowed to react with the room temperature liquid. 1,4-DiphenyIbutane (Aldrich) was allowed to react in the same manner. (c) Synthesis of the Phenyl Ether of Methyltetraethylene Glycol. The compound was prepared by tosylating the monomethyl ether of tetraethylene glycol (Aldrich) and reacting the tosylate with phenolate anion as described above for the synthesis of apdiphenylhexaethylene glycol. (i) CH30(CH2CH20).,Ts. NMR (acetone-d,, 60 MHz): 6 2.40 (br s, 3 H, ArCH3), 3.23 (s, 3 H, CH30-), 3.45-4.35 (m, 16 H, -OCH2CH20-), 7.30 and 7.72 (AA’BB’, 4 H, JAB= 8 Hz, ArH). (ii) CH30(CH2CH20)4C6H5. NMR (acetone-& 60 MHz): 6 3.23 (s, 3 H, ArCH,), 3.42-4.23 (m, 16 H, -OCH2CH20-), 6.70-7.42 (m, 5 H, ArH). Reactions of this liquid with vanadium and molybdenum atoms are described later in this paper.

Results and Discussion To gain an appreciation of the earlier work and that which follows, we present a brief review, with some supplementary new results of data regarding Mo/DC 510 reactions obtained by the microscale metal vapor technique. This information forms a necessary basis for the subsequent presentation. The existence of (arene)2Mo complexes in DC 510 can be considered e~tablished.~.’Figure 1 shows the UV-vis absorption spectra (new data) of ( C 6 H 5 C H 3 ) 2 Min~pentane ( E = 1.1 X lo5, ,A, = 318 nm) (Figure 1A) and in poly(dimethylsi1oxane) (DC 200) (Figure 1B) and for the complex prepared in the DC 510 polymer. In supplementary work we have synthesized these compounds on the gram scale by electron-beam evaporation of Mo accompanied by reaction with toluene or DC 510.15 The major features of importance in the figure are the intense MLCT absorption centered at 318 nm and a weak absorption at 395 nm. On the basis of the measured peak heights, the ratio of the 318/395-nm transitions is 80:l for the (arene)2Mo complex regardless of the solvent medium (pentane or polymer). In Figure 2 we represent graphically the results of depositing increments of Mo into a thin (10-100 pm) 250 K film of DC 510. The initial Mo deposition represented in Figure 2 produces (arene)2Mo. The absorption at 3 18 nm, by comparison with Figure 1, is characteristic of the MLCT transition of the molybdenum sandwich complex. Increasing the metal loading (at approximately (15) Ozin, G. A,; Andrews, M. P. Znorg. Chem., in press.

*

2590 Inorganic Chemistry, Vol. 25, No. 15, 1986

Andrews and Ozin

290 K

,

--

-_I

- 40 f

1’” 20

10 20 Y

8

16

24 ,ug V

32

40

1

48

8

1

I

I

-

I

I

1

40

1

I 72

I

I

I

pug varldkm

Deposrted

Figure 5. (A) Graphical representation of growth characteristics: (0)growth of the (arene)*Vcomplex when Santovac-5 is titrated with,vanadium atoms; ( 0 )growth of the second species, which is associated with a vanadium dimer. The peak maximum in the absorption spectrum is plotted as a function of the amount of deposited metal in each case. (B)Graphical representation as in part A above, except that the reaction occurs between V and a,w-diphenylhexaethyleneoxide. The effect of V deposition rate is shown by comparing the shape of the two growth curves.

2X mmol h-I deposition rate) to roughly mol results in the growth of a new absorption at 418 nm (Figure 2). As the amount of molybdenum in the film is gradually increased, more transitions can be observed at 502, 578, and 640 nm (Figure 2). The variation of the absorbance for each band as a function of metal loading is depicted in Figure 3 and confirms that each bund represents a separate species as shown by their independent growth characteristics. The following extinction coefficients were estimated by using the data of Figure 2: c318 = 2.7 X lo5, tdI8= 7.4 X lo4, e502 = 2.2 X lo4 L mol-l cm-’. The value for the polymer supported (arene)zMo complex is within a factor of 2 of the value given above for the metal-to-ligand charge-transfer absorption of bis(to1uene)molybdenum in pentane.’ The order of stability of the species giving rise to those bands red-shifted from the 318-nm MLCT absorption is the same as the order of production. Except for the absorption associated with the (arene)2Mocomplex, all other absorptions gradually disappear on warming of the polymer medium to 300 K. Furthermore, these latter absorptions are rapidly quenched on allowing r l torr of O2to diffuse into the fluid at 290 K. In contrast, the polymer bound bis(arene)molybdenum(0) complex is oxidized only slowly to the bis(arene)molybdenum(I) analogue.’ The above data with minor modifications are essentially representative of our findings for the Ti, V, and Cr reactions with DC 510 as well as their bimetallic

(iv) Bimetallic M/M‘/DC 5 10 studies provide additional support for the metal nuclearity and metal-localized electronic

(16) Ozin, G. A. Faraday Symp. Chem. SOC.1980,14,7 and referenca cited

(v) Microsolution spectroscopy experiments on V and Cr atom titrations of bis(ta1uene)vanadium and bis(to1uene)chromium dissolved in 140 K methylcyclohexane liquid films indicated the arene stoichiometry might be two for the binuclear species.I9 (vi) Simple series parallel competitive kinetic schemes were devised that account for the cluster growth curves for Vn-,Crn-, and Mo,/DC 510, up to the n = 4 stage for Mo (see part 2 of this series of papers). Reactions of Transition-Metal Atoms with Arenes and AreneFunctionalized Alkanes and Oligo(ethy1ene oxides). In order to examine in more detail the origin of the metal concentrationdependent absorptions in DC 510 we developed the synthetic strategy sketched in Figure 4. From top to bottom the scheme shows metal vapor syntheses in matrices varying in temperature from 293 to 12 K. In common with the DC 510 polymer the benzene ring is repeated from ligand to ligand. A subset of ligands also contains ether linkages. Considering the exemplary vanadium atom, quantitative vaporization into neat liquids or solutions in every case (A-F) resulted in the production of the bis(arene)vanadium complex recognizable by its characteristic chargetransfer absorption at 330-340 nm (depending on the ligand) and its EPR spectrum. For the a p d i p h e n y l compounds, bis(arene)metal complex formation is possibly accompanied by oligomerization leading to chain lengthening. If we trace routes A and B in Figure 4 for vanadium atom reactions, in each case a band is produced at ~ 4 5 nm 5 in addition to the intense MLCT absorption of the sandwich complex. As found for reactions of this transition metal with DC 510, the species absorbing a t 455 nm shows growth characteristics independent of the bis(arene)vanadium complex$*5although a rate dependence shows up at the slowest depositions (see part 2 of this series of papers). New examples of this behavior are shown in Figure 5. They can be compared with those of vanadium deposited into DC 510 at 250 K.4,5 Since reactions of vanadium with apditoluyldecane or 1,4-diphenylbutane lead to the development of the 455-nm absorption, we conclude that ether oxygens are not implicated in its production. Clearly the ability to establish an initial chain

(17) Gilbert, J. R.;Leach, W. P.; Miller, J. P. J . Oraanomet. Chem. 1973. 56, 295. (18) Andrews, M. P.; Ozin, G. A. J . Phys. Chem. 1986, 90, 1245

(19) Ozin, G. A.; Andrews, M. Angew Chem., Suppl. 1982, 381.

counterpart^.^-^ Arguing heuristically for the existence of polymer-supported metal clusters, we note the following: (i) The excitation energies of the multimers M, ( n 2 2) in solid rare-gas matrices red shift with increasing nuclearity.16 (ii) The gas phase secondary ions [(arene)zCr2(C0)3]+,[(arene),Cr3(C0),]+, and [(arene),Cr,]’ are known to exist.” (iii) Concentration studies of the cocondensation of V, Cr, and Mo with C6H6:Ar = 1:10 at 12 K reveal bands on the low-energy side of the (&C6H6),M MLCT band, possessing a metal concentration dependence consistent with that expected for binuclear metal species. These lower energy transitions are assigned to metal-localized e x c i t a t i o n ~ . ~ - ~ J ~

therein.

I

Inorganic Chemistry, Vol. 25, No. 15, 1986 2591

Reactions of Transition-Metal Atoms with Polysiloxanes

@--?

v

Table I. Results of Optical Absorption Spectroscopy of Microsolution or Matrix-Isolation Reactions of Vanadium Atoms with Various Functionalized Arenes abs max, nm reacn substrate (arene)2V ‘V,” temp, K 340 458 290 Santovac-5 a,w-diphenylhexaethyleneoxide 335 454 290 a,w-ditoluyldecane 340 460 290 phenylhexaethylene oxide 330 452 280 monomethyl ether DC 510 323 455 250 1,4-diphenyIbutane. 345 455 245 phenyldecane 338 455 225 benzene (toluene) 328 (335) 450 (450) 77

225K

328 (335) 450 (450) 328 450 478, 555-590, 740-840

benzene:argon 1:10 argon or alkanes

12 12 12

I

P

250

330

4x)m

Figure 6. Microsolution reaction of vanadium atoms with 1-phenyl-ndecane at 255 K: spectrum A, base line; spectrum B, = 3 X mmol of vanadium deposited at 2 X lo4 mmol h-I; spectrum C, recorded at 300 K after removing unreacted ligand in vacuo. propagation step via bis(arene)metal complex formation is still possible through sequences A X . One can imagine that a medium of increasing viscosity due to increasing molecular weight might act in some physical manner to immobilize small transition-metal cluster molecules. The view here is that of a trapping medium comprising randomly interpenetrating molecules oligomerized as a result of the formation of bis(arene)metal complexes. Experiments were undertaken in which vanadium atoms were vaporized into a variety of high molecular weight poly(dimethylsi1oxanes) (M,104-106). The optical absorption spectrum evidenced only an unstructured absorption rising smoothly toward the ultraviolet. This is typical behavior for absorption and scattering by colloidal metal. Arguing against the necessity of having an oligomeric or polymeric medium to stabilize the species absorbing at 455 nm, we have found that reactions of vanadium atoms with 1phenyl-n-decane, which is liquid at 255 K, rapidly gave rise to the 455-nm band. This reaction is important for another reason. Figure 6B shows the result of depositing -3 X mmol of V (deposition rate = 2 X lo4 mmol h-l) into a thin film of phenyldecane maintained at 255 K on a 3-cm2 quartz window. By a careful warming of the fluid to 260 K, the unreacted phenyldecane could be removed under dynamic vacuum. Spectrum C was recorded at 300 K and shows the metal-ligand charge transfer and the 455-nm absorptions coexisting. The rising baseline between 350 and 250 nm is due to scattering by the inhomogeneous residual organometallic film, interspersed with minute gray particles, which were probably colloidal vanadium. Unlike the benzene or toluene analogues, the products of the reaction of vanadium with phenyldecane are involatile under 10-6 torr, at room temperature. A weak spectrum showing both absorption maxima was obtainable after 13 h a t 300 K. On eliminating all substituents linked to the phenyl ring (Figure 4) we return to the metal concentration studies involving V C,&/Ar where a dimeric vanadium compound was first inferred. It therefore appears likely that the species responsible for the 455-nm absorption in V/DC 510 and in the phenyl rich systems of Figure 4 can be assigned to a vanadium dimer molecule interacting in an as yet unspecified manner with the phenyl substituent(s). The results of optical absorption spectroscopy of the series of microsolution and matrix-isolation reactions represented in the scheme of Figure 4 are summarized in Table I. The position

i,

I

xo nln Figure 7. UV-visible absorption spectroscopic study of the reaction of Mo atoms with a,w-diphenylhexaethyleneoxide at 290 K: A, pure oligomer; B-D, increasing Mo atom concentration. xo

xx,

400

of the MLCT absorption due to (arene)2V varies from 328 to 345 nm, showing a red shift as the ring is substituted with alkyl or ether groups and as the surrounding medium is changed from aliphatic/arene to ether/arene. The uV2nabsorption is not uniformly red-shifted, showing an average value of 455 f 5 nm and suggesting excitations that are mainly metal-localized. Neither the MLCT nor the “Vzn band shows a substantial variation in the position of the peak maximum as the temperature is varied to 12 K. R e a c t i o ~of Arenes with Mo Atoms. Returning to the scheme shown in Figure 4, a similar set of reactions was also carried out for Mo, with similar findings. Reactions with Mo gave rise to a MLCT absorption varying from 323 nm in DC 510 to 327 nm in 1,lo-ditoluyldecane. A second absorption showing the familiar metal atom concentration dependence was detected a t 4 1 8 4 2 2 nm in every case. A typical series of room-temperature microscale reactions between Mo atoms and the seven-oxygen a,o-diphenylhexaethylene oxide is shown in Figure 7. This spectrum clearly displays three of the growth states encountered for reactions of Mo with DC 510 (Figure 2). Regarding the ether oligomer, the absorptions at 418 and 502 nm coincide with those detected in DC 510 fluid. Additional reasons for choosing the oligo(ethylene oxides) derive from their proven ability to complex metal

Andrews and Ozin

2592 Inorganic Chemistry, Vol. 25, No. 15, 1986

250 K

418nm

n

/

250

300

350

400

450 nm

Figure 8. Results of Mo atom titration of M (C6H6)2Mo dissolved in DC 200 at 250 K: curve A, base line; curve B, 0.5 X lo4 mmol mmol (corrected) of Mo deposited; curve C, an additional 0.3 X (corrected) of Mo deposited. Curves D-G represent the outcome of warming the matrix through various temperature increments.

ions, an ability that might extend to a metal atom. To the results of the above series of reactions we can add the outcome of depositing Mo atoms into DC 200 containing dissolved bis(to1uene)molybdenum. This liquid poly(dimethylsi1oxane) was chosen to mimic some of the properties of DC 5 10 (similar viscosity and molecular weight). In this reaction matrix, the sandwich complexes are not bound to the polymer chain, but the situation provides a polymeric solvent medium similar to DC 5 10 and loosely represents the case in which all available phenyl groups have been converted to complexes. Figure 8 lucidly expresses the results of this experiment. Curve A represents 10-4 M bis(to1uene)molybdenum dissolved in DC 200 (60 MLof fluid spread over an area of 3 cm2) at 250 K in vacuo. Curves B and C represent successive increments (0.5 X lo4 and 0.3 X IO4 mmol (corrected)) of Mo mmol h-' into this film. atoms deposited at a rate of 2.4 X The consumption of the bis(to1uene)molybdenum complex and the simultaneous growth of the absorption at 418 nm are manifest. Warming the film to 270 K (curve E) and then to 295 K (curves F and G ) appears to result in the preferential loss of residual bis(to1uene)molybdenum. There is also a simultaneous but less rapid decrease in the intensity of the 418-nm band. A subsequent experiment established that the (C6HSCHJ2Mocomplex gradually torr). This metal sublimes from the DC 200 film at 290 K (vapor experiment suggests the possibility of generating and separating the thermally unstable reaction product from the starting sandwich complex by preferential sublimation and hence provides an opportunity for obtaining a stable product for a detailed spectroscopic and/or X-ray study on this fascinating class of (arene)metal cluster complexes. Hopefully this goal will be achieved in the near future. Interestingly, the polymer medium seems to impede the thermally activated decomposition of the species responsible for the 41 8-nm absorption. Reasoning from the results of our studies of vanadium and chromium, we suggest that a molybdenum dimer is the likely candidate for the origin of this band. We have carried out a molybdenum/benzene concentration study and a kinetic analysis of the type described

earlier.'* A plot of log (A418/A320) vs. log Ro for the species absorbing at 320 and 418 nm gave a straight line with a slope of 0.95, implicating a molybdenum dimer as the carrier of the 418-nm absorption. In the MO/DC~OO/(CH~C~H,)~MO thin-film experiments there was never any evidence for additional absorption(s) to the red of the 418-nm transition analogous to that observed under high Mo loadings in reactions of Mo with the a,w-diphenyloligo(ethy1ene oxides) or to those observed in reactions with the DC 510 polymer. An interpretation of this observation is presented later in the paper. EPR Experiments. A series of microscale EPR experiments on the V/DC 510 system indicated only the eight-line hyperfine spectrum due to the polymer-supported DC 5 10/(arene)2V;zo careful examination under various conditions of vanadium loading and temperature (293-12 K) disclosed only that the bis(arene)vanadium complex was EPR active under our conditions. The same behavior was found in this study when V was deposited into 1,lO-diphenylhexaethyleneglycol or cocondensed with benzene at 77 K at high V atom concentrations. Moreover, when Cr atoms were deposited into bis(to1uene)chromium (EPR silent) dissolved in DC 200 (250 K microscale) no EPR signal developed. In the reverse reaction, Cr atoms were deposited into bis(to1uene)vanadium dissolved in DC 200. The eight-line spectrum decreased in intensity with the amount of deposited chromium, showing that reaction had occurred, but there was no evidence of a new EPR signal either at 250 or 12 K. IR and Raman Experiments. The room-temperature infrared spectrum of DC 5 10-bound bis(arene)chromium is reported to show absorptions at 488 (v(MR)) and 458 cm-' (asymmetric ring tilt).3 This spectrum is characteristic of a neat thin film of the polymer prepared by the rotary-resistive metal vapor method. The UV-visible absorption spectrum of the same polymer is free of bands to low energy beyond 395 nm. Bis(benzene)chromium exhibits absorptions at 492 (A2,, v(MR)) and 459 em-, (Elu asymmetric ring tilt) in KBr. In our own experiments we were unable to detect the IR absorptions a t 488 and 458 cm-I in DC 510/Cr because the low-energy region is congested with polymer absorptions that obscure the 400-500-cm-' window. A band has however, been detected at 432 cm-' in Santovac 5/Cr (a liquid pentaphenyl ether) prepared in situ in a microscale IR experiment. Similarly the metal-ring modes of DC 510/V expected in this region are obliterated by the polymer absorptions. In contrast the low-frequency Raman bands of the metal atom free polymer, although rather broad, are few in number and occur at 490 cm-' (polarized) and 180 cm-I (depolarized). Since the Raman experiment can provide evidence of a symmetric M-M stretching mode2, in a dimeric species of the type we have postulated, a series of matrix isolation and microsolution Raman experiments was initiated. Depicted in Figure 9 is the Raman spectrum resulting from depositing C r atoms into DC 510 at 240 K. The spectrum was recorded of the sample cooled to 100 K, under conditions of low Cr loading such that only the bis(arene)chromium complex is presumably present. The upper and lower traces respectively show the perpendicular and parallel polarizations. The symmetric metal ring stretching mode can be. assigned with some certainty to the band at 290 cm-' on the basis of the A,, (polarized) vibration observed at 277 cm-I in (C6H&Cr.22 The remaining bands arising a t 350, 424, and 452 cm-' for this polymer-supported complex are assigned on the basis of analogous modes that occur at 335 cm-' (El, (symmetric tilt)) and 459 cm-' (Elu(asymmetric tilt), IR active), in the more symmetric bis(ben2ene)chromium complex. The v(MR) that occurs at 490 cm-' in the infrared spectra of both the polymer-bound sandwich complex and the (Azu) bis(q6-benzene)chromium complex is quite possibly obscured by the intense 490-cm-I polymer band. The v(MR,) assignment at (20) Andrews, M. P.; Mattar, S. M.; Ozin, G. A. J. Phys. Chem. 1986, 90, 1037.

(21) Shriver, D. F.; Cooper C. V., 111. Ado. Infrared Raman Spectrosc. 1980, 6, 127. (22) Fritz, H. P. Adu. Organomet. Chem. 1964, 1 , 239

Reactions of Transition-Metal Atoms with Polysiloxanes

Inorganic Chemistry, Vol. 25, No. 15, 1986 2593 Table 11. Reactivity of Transition Metal Atomic Vapors toward Liquid Polymeric and Oligomeric Materials“ 1. DC 510 + M(g) ‘HNMR-only a multiplet for the bis(arene)metal complex at 6 5.39 (Ti), 4.85 (Mo), 4.59 (Cr); infrared shows no Si-0-M for M = Ti, V, Cr, Mo 2. DC 510 + Ti(g) (DC 510)[(arene)2Ti]% Ti02 + DC 510 GPC=> shows molecular weight unchanged*’ 3. DC 200 (poly(dimethylsi1oxane)) or paraffin oil + M(g) colloidal metal

-

0--H

co Lo

P o

6.

CH3O

7.

CIOHL~OCHI

~

O

t Cr(g)

-

C t ~~ ( H g )

C r ( I I I ) alkoxide

~ 1arrnr)zM complex

see 1 above, NMR

data

( I 1 77 K

t Crlg)

,*) ~ C 0 n 4 * n ’ ’

CloHaOCH3 t Crlcolloid)

r e f 24

‘For a list of solvents that are “inert” toward metal atoms under normal synthetic conditions see: Blackborrow, J.; Young, D. Metal Vapour Synihesis in Organometallic Chemistry; Springer-Verlag: New York, 1979.

502

400

300

200

1 0 3 cr

Figure 9. Raman spectra representing the outcome of metal atom titrations into the DC 510 polymer and matrix-isolation reactions with toluene/vanadium. The upper trace (recorded at 100 K, products prepared from reaction with Cr atoms at 240 K) shows Raman-active modes (polarized) of polymer-supported (arene)*Cr. The bottom trace shows the outcome of depositing 3 X lo-’ mmol of vanadium at 1.6 X mmol h-’ into 240 K DC 510. Recorded at 100 K, the spectrum shows a band at 155 cm-l (strongly polarized), which is assigned to the vanadium dimer and is not found in the Raman spectrum of bis(to1uene)vanadium (details inserted).

424 cm-I in the Cr/DC 510 spectrum is tentative. mmol h-’ into DC When vanadium is deposited at 1.6 X mmol of vanadium, 510 a t 240 K to a total loading of 3 X a transparent deep yellow-gold product develops. Under these deposition conditions the vanadium dimer species is expected to form readily. New Raman bands at 520,430,270, and 155 cm-’ appear, all of which are polarized by comparison with the depolarized 180-cm-’ polymer reference line (Figure 9). The bands at 520, 430, and 270 cm-I, by analogy with the vibrational spectrum of (C6H&V and other discrete bis(arene)metal sandwich complexes,zz are assigned to M-arene ring deformation and stretching modes. The band at 155 cm-l has no counterpart in bis(arene)vanadium generated in situ in a V/toluene 77 K matrix or in reactions of V with DC 510 under mononuclear metal atom conditions and therefore may be characteristic of the polymersupported divanadium species. Low-frequency Raman spectra of other M-M-bonded dimer and cluster complexes range typically from 100 to 425 cm-I, depending on the multiplicity of the bonding and the nature of the bonding interactions with associated ligands.zl Within a given class of dimers the stretching frequencies can be spread over a broad range. Thus the formally quadruply bonded Mo-Mo dimers exhibit vibrational frequencies from 339 to 413 cm-I, although the ran e of Mo-Mo internuclear distances is rather small (2.05-2.13 ). These observations should be contrasted with those of analogous dirhenium complexes. These compounds show a very narrow range of stretching frequencies, 272-293 cm-I. It is still largely a matter of conjecture whether the stretching frequencies reflect trends in metal-metal bonding or the influence of coupling of the symmetric metal-metal stretching mode to M-ligand displacement coordinates. Thus the Raman band we have observed at 155 cm-I (polarized) in the high-loading V/DC 510 system may be associated with the sym-

1

metric V-V stretching mode coupled with the arene ligands of DC 510. Reactions of Organic and Inorganic Molecules with Metal Vapors. Relevance to Metal Atom Multimerization in AreneContaining Media. There have been a number of applications of metal vapors to effect functional group transformations of molecules. These applications include (i) the radical reduction of cyclohexanone to [ 1,l’-bicyclohexyl]-1 ,l’-diol by early-transition-metal atoms and clusters23and (ii) deoxygenation and des u l f ~ r i z a t i o nof~ ~a host of organic compounds. With the observation that the deposition of nickelz5or zirconium26vapor into various organic matrices can result in cleavage of the organic species, an explanation of the red-shifted bands seen in absorption for the M/DC 510 and M/arene reactions must include reference to the possibility that metal atom induced fragmentation reactions leading to organometallic species might be responsible for the additional absorptions. We have examined this latter possibility by performing a series of control experiments on the same and related compounds, the results of which are summarized in Table 11. Except for the N M R experiments, the conversion of Ti/DC 5 10 to TiOz and analysis of the liberated DC 510 by gel permeation chromatography (GPC)?’ all reactions could be carried out as both microsolution or gram-scale rotary reactor preparations. In general, we found that silyl ethers and alkyl ethers are unreactive toward the metal atoms examined. A similar conclusion regarding alkyl ethers was reached by G l a d y s ~ .Titanium ~~ atoms, which are particularly o ~ o p h i l ishowed ~ , ~ ~ no ~ ~tendency ~ to cleave the backbone S i 0 linkages of the DC 510 chain (GPC) under the usual rotary reactor metal vapor conditions (reaction 2). Depositions of Ti, Cr, and Mo into DC 510 gave room-temperature ‘H NMR multiplets only of the respective bis(arene)M products at 6 5.39 (Ti), 4.59 (Cr), and 4.85 (Mo) in agreement with previous observation^.^ There was no evidence of a band in the infrared region 910-930 cm-’, diagnostic of an Si-&Ti linkage,z8 in agreement with the GPC analysis. Negative IR evidence regarding Si-0-M resulted also (23) Miller, J. T.; DeKock, C. W. J . Org. Chem. 1981, 46, 516. (24) Togashi, S.; Fulcher, J. G.; Cho, B. R.; Hasegawa, M.; Gladysz, J. A. J . Org. Chem. 1980, 45, 3044 and references therein. (25) Matsuo, K.; Klabunde, K. J. J . Org. Chem. 1982, 47, 843. Klabunde, K. J.; Tanaka, Y. J. Mol. Catal. 1983, 21, 57. (26) Remick, R. J.; Asunta, T. A.; Skell, P. S.J . Am. Chem. SOC.1979,101, 1320. (27) We acknowledge the technical assistance of Dr. Colin Francis in car-

rying out this experiment. (28) Borisov, S. N.; Voronokov, M. G.; Lukevits, E. Ya. Organosilicon Heteropolymers and Heterorompounds; Plenum: New York, 1970.

2594 Inorganic Chemistry, Vol. 25, No. 15, 1986 for M = V, Cr, and Mo reactions with DC 510. As remarked upon earlier and described in ref 4-8, vaporization into poly(dimethylsiloxane) or paraffin oil resulted only in dispersions of colloidal metal (V,Cr,Mo). In the microsolution experiments involving vaporization into DC 200 there was no evidence of UV-visible absorptions of the type observed when the metals were deposited into the DC 510 polymer. The same results were obtained for reactions of the dimethyl ether of tetraethylene glycol with vanadium or chromium. By contrast, Cr atoms reacted with the acidic hydroxyl proton of the monomethyl ether of the aforementioned molecule to give a green product, which is probably the chromium alk~xide.,~Unlike the Cr"'(OMe), this complex is soluble in THF. The infrared spectrum measured in mulls and KBr disks shows a strong band at 1053 cm-' due to v(C-O)M, and a band at 600 cm-l ( ~ 4 M - 0 ) ) The ~ ~ ~electronic spectrum was measured in situ (microsolution) for depositions of C r into the monomethyl ether at 260-290 K; however, no transitions due to the expected alkoxide product were observed under the experimental conditions (8 X lo4 mmol of Cr; 7 X lo4 mmol h-l). A gray-green product was isolated from the reaction of Cr with the monomethyl ether in a rotary metal vapor reaction and gave an absorption spectrum that showed three transitions: an intense absorption at 43 500 cm-', and two much weaker bands at 24000 and 17 240 cm-I. This absorption pattern is similar to that of known chromium(II1) alkoxides,2" for example Cr(OCH,), (E, = 47 200 cm-' (charge transfer); E2 = 24 200 cm-l; E3 = 17000 cm-I). Depositions of vanadium atoms into the monomethyl ether at room temperature (2.8 X loT3mmol of V a t 2 X lo-, mmol h-l), or poly(ethy1ene oxide diol) (average mol wt 400) (1.7 mmol h-l) gave no EPR signal, although mmol of V; 2 X a resonance may be expected.29b Gladysz has d e m ~ n s t r a t e dthat ~ ~ decyl methyl ether does not react with chromium atoms on warming the 77 K cocondensate to room temperature. In fact, it is the relative inertness of ether and alkane compounds under most metal vapor synthetic conditions that makes them desirable solvents for metal atom reactions. Insertion products of the type described in K l a b ~ n d e and ,~ Webb et aL30appear most likely to occur on the surface of small metal aggregates, and the production and reactivity of these too are sensitive to the conditions of sample preparation. Our own experience has shown that very rapid electron-beam vaporization of vanadium in a 77 K cocondensation experiment involving toluene in hexane produces a highly reactive dispersion of colloidal metal in the frozen matrix, in addition to the bis(arene)vanadium complex, whose presence was inferred from the color of the codeposit. On meltdown in vacuo, rapid bubbling was observed and the bis(to1uene)vanadium that had probably formed on deposition was completely destroyed. Bibenzyl was separated as one of the products of this reaction. The remaining products appeared to be a complex series of aryl-substituted alkanes and olefins from the proton N M R . In conclusion we regard it as unlikely that organometallic fragments resulting from cleavage of alkyl or ether bonds are responsible for the trends in the optical spectra observed in reactions of transition-metal atoms with the arene containing systems presented in this paper and earlier reports. Mass Spectral Detection of Organometallic Metal Dimers and Heterodimers and Higher Multimers. Support for the existence of compounds formulated as (C6H6)2M2comes from the gas-phase ion-molecule reactions involving C&Cr( CO)3 reported by Gilbert et In Gilbert's M S experiments the existence of [(C6H6),Cr,]+ was established. This dimer species is isoelectronic with our proposed (C6H6)2VCrheterodimer. The same heterodimer, ligated with CO, has been synthesized in the gas phase by pulsed-laser volatilization and source ionization of vanadium in the reaction3I (29) (a) Brown, D. A.; Cunningham, D.; Glass, W. K. Chem. Commun. 1966, 306. (b) Bradley, D. C.; Mehrotra, R. C.; Gaur, D. P. Metal Alkoxides; Academic: New York, 1978. (30) Reid, A. H., Jr.; Shevlin, P. B.; Yun, S.S.;Webb, T.R. J. Am. Chem. SOC.1981, 103, 709. (31) Kappes, M. M.; Staley, R. H. J. Chem. Phys. 1982, 86, 1332.

Andrews and Ozin

v+ t

CdC016

/

\

VCr(C0):

(60%) t CO

VCr(C0):

(40%)

t ~ C O

We note in addition that the heterodimer, M I I C ~ ( C O ) ~(80% + yield), is isoelectronic with our proposed (arene),Cr2 compound. The above ion-molecule reaction between V+ and Cr(C0)6 is particularly significant because the observed products are the result of a first-step direct addition of the metal cation to a metal center in Cr(C0)6, a process in line with our proposed view of direct addition of a metal atom to a preformed bis(arene)metal site.I9 In this context, two other very recent gas-phase studies are worthy of mention. First, Hettich and Freiser3, have shown that gaseous VFe+ (isoelectronic with Cr,) rapidly dehydrogenates cyclohexene according to VF;

t

0-

VFe(C6H6?

+

2H2

forming VFe(C6H6)+containing a metal-metal bond. In a secondary reaction this species dehydrogenates another molecule of cyclohexene to give a 45% yield of VFe(C6H6),+ through the reaction

't

VFe(CsH6)

0

Vh(CsH6):

+

2H2

A bis(benzene)dimetal structure similar to that of our proposed M,(arer~e)~ clusters was suggested for VFe(C6H6),+. Second, recent gas-phase studies of Whetten et have revealed that a beam of Pt, clusters react with benzene to form dibenzene adducts of the form Ptn(C6H6)2, where n = 2-6. No structural information was reported in these studies although the stoichiometric resemblance to our proposed M,(arene), clusters is striking. Summary In earlier studies from this laboratory, single metal atoms have been found to react with one or two benzene rings in rare-gas matrices to form the respective half-sandwich (v6-C6H6)Mor bis(benzene)metal complex (v6-C6H6),M, depending on the re]ative concentrations of the various components.34 At higher metal atom concentrations, matrix kinetic studies have revealed the existence of (C6H6),M2 for M = v , Cr, and Mo where n may be 2.4-73'8 With the use of a poly(dimethylsi1oxane-co-methylphenylsiloxane) liquid, the metal concentration dependence of the changes in the absorption spectrum of the products has been followed near room temperature. These studies were made possible by the development of an in situ microscale metal vapor spectroscopic technique, which provided controlled conditions under which we probed the outcome of metal vapor liquid and solution phase reaction^.^ The technique was shown to form a natural bridge between matrix isolation and macroscale spectroscopic and synthetic method^.^ For example, at sufficiently high loadings of vanadium atoms in DC 5 10, a new band appeared at 455 nm, which showed a concentration dependence independent of that shown by the polymer-supported bis(arene)vanadium c ~ m p l e x . ~ - ~ ~ Similar results were obtained for Ti, Cr, and M O . A~ synthetic strategy was worked out in which it was shown that the only condition necessary for the Occurrence of the new species was the presence of arene ligands. Thus, the outcome of metal atom deposition into low-temperature, methylcyclohexane solutions of (arene)2M, (M = Cr, V), suggested that a metal dimer species forms by direct addition of the atom to a preformed parent (aret~e)~M complex. The arene stoichiometry suggested was tw0.I9 In the present study, some further insight into the mode of stabilization of metal clusters in the DC 510 system is obtained from similar studies involving different phenyl-substituted mo(32) Hettich, R. L.; Freiser, B. S . J. Am. Chem. SOC.1985, 107, 6222. (33) Whetten, R. L.; Cox, D. M.; Trevor, D. J.; Kaldor, A. Surf.Sci. 1985, 156, 8-35 and references cited therein. (34) Ozin, G. A,; Mattar, S.;Huber,H.; Andrews, M. P.; McIntosh, D. F. J. Am. Chem. SOC.1983, 105, 6170.

Inorg. Chem. 1986, 25, 2595-2599 nomers and oligomers in metal vapor reactions surveying a thermal gradient varying from 12 to 300 K. In the context of the M,/DC 510 work, a series of control experiments were performed which demonstrated that metal atoms are unreactive toward substituents (ether linkages, Si-0, C-H, etc.) other than the arene. In addition an experiment was designed which demonstrated that molybdenum atoms can be used to quantitatively titrate (arene)*Mo dissolved in an unreactive liquid polymer, exemplified by poly(dimethy1siloxane), DC 200. The product in DC 200 absorbed a t 418 nm, identical in position with the Mo/DC 510 concentration-dependent absorption, thereby providing further credence for the binuclear formulation (arene),Mo,. Acknowledgment. G.A.O. and M.P.A. gratefully acknowledge the generous financial support from the NSERC Operating and

2595

Strategic Grants Programmes and the Connaught Foundation of the University of Toronto. The award of a 3M Corp. grant to G.A.O. is also greatly appreciated. Registry No. HO(CH2CH20)6H,261 5-15-8; HO(CH2CH20)2H, 11 1-46-6; BrCH2CH20CH2CH2Br,5414-19-7; T S O ( C H ~ C H ~ O ) ~ T S , 42749-27-9; TSCI, 98-59-9; C6H5O(CH2CH20)6C6H5, 20740-88-9; C~HSOH,108-95-2; C ~ H S O ( C H ~ C H ~ O ) ~20768-77-8; C ~ H ~ , C6HSO(CH ~ C H ~ O ) S C ~20612-83-3; H~, CHPC~H~O(CH~CH~O)~C~H~CH~, 102073-92-7; C ~ H S O ( C H & H ~ O ) ~ C20740-89-0; ~H~, CH,-pC6H4(CH2)loC6H4-pCH3,102073-91-6; p-BrC6H4CH3, 106-38-7; Br(CH2)loBr,4 101-68-2; CH,0(CH2CH20)4Ts,6292 1-76-0;CH30(CH2CH20)4C6H5, 80392-31-0; V, 7440-62-2; Cr, 7440-47-3; Mo, 7439-98-7; CHsO(CH2CH20)4H, 23783-42-8; CH,O(CH2CH20)4CH,, 143-24-8; CIOH210CH3, 7289-52-3; Santovac-5, 2455-71-2; 1,3-diphenylbutane, 1083-56-3;phenyldecane, 104-72-3; benzene, 71-43-2; toluene, 108-88-3.

Contribution from the Department of Chemistry, University of California, Irvine, California 92717

Bis(nitroxy1) Adducts of Bis(bexafluoroacetylacetonato)manganese( 11). Preparation, Structures, and Magnetic Properties Michael H. Dickman, Leigh C. Porter, and Robert J. Doedens* Received January 14, 1986 Adducts of bis(hexafluoroacetylacetonato)manganese(II) with the cyclic nitroxyl radicals 2,2,6,6-tetramethylpiperidinyl-1-oxy (tempo) and 2,2,5,5-tetramethylpyrrolidinyl-1-oxy (proxyl) have been prepared. Both products are bis adducts, Mn(F6acac)2L2 (L = tempo, proxyl), and have been shown by crystal structure analyses to have centrosymmetric trans octahedral structures. In each case, the metal ion is bound to two chelating F6acac ligands and two monodentate nitroxyls, with Mn-O distances ranging from 2.1 2 to 2.16 A. The configurations of the metal-bound nitroxyls closely resemble those of analogous uncoordinated radicals. Magnetic susceptibility measurements (6-300 K) are consistent with an S = 3/2 ground state for both complexes, with a significant population at higher temperatures of one or more states of greater multiplicity. The susceptibility data are well-represented by interaction model, with J(Mn-nitroxyl) values of -79 (tempo adduct) and -105 cm-' (proxyl adduct). a spin-only (1/2-5/2-'/2) Both adducts cr stallize in the monoclinic space group P2'/c with Z = 2. For the tempo adduct [Mn(C5HF602)2(C9H18NO)2], a = 10.698 (5) b = 14.715 (6) A, c = 12.378 (5) A, and j3 = 113.48 (3)'. For the proxyl adduct [Mn(C5HF602)2(C8H16N0)2], a = 10.472 (2) A, b = 14.488 (4) A, c = 12.090 (2) A, and j3 = 111.08 (2)O.

1,

Introduction Complexes in which a transition metal center is coordinated to a nitroxyl radical function have been the subjects of considerable recent interest in our laboratory and e1se~here.I-l~One of the motivations for the study of these compounds is the potential they give for the investigation of interactions between the free spins on the nitroxyl radical and on a paramagnetic metal center. Compounds prepared and characterized to date have shown interactions spanning a wide range of magnitudes. Strong inter(1) Richman, R. M.; Kuechler, T. C.; Tanner, S.P.; Drago, R. S.J . Am. Chem. SOC.1977, 99, 1055-1058. (2) . , Draao. R. S.;Kuechler. T. C.; Kroeger, - M. Inorf. - Chem. 1979, 18, mi-2342. (3) Anderson, 0. P.; Kuechler, T. C. Inorg. Chem. 1980, 19, 1417-1422. (4) Dickman, M. H.; Doedens, R. J. Inorg. Chem. 1981,20,2677-2681 and references therein. (5) Grand, A.; Rey, P.;Subra, R. Inorg. Chem. 1983, 22, 391-394. (6) Porter, L. C.; Dickman, M. H.; Doedens, R. J. Inorg. Chem. 1983,22, 1962-1964. (7) Bcnelli, C.; Gattexhi, D.; Zanchini, C. Inorg. Chem. 1984,23,798-800. (8) Sharrock, P.; Melnik, M. Abstracts of Papers, 23rd International Conference on Coordination Chemistry, Boulder, CO; Cooperative Institute for Research in Environmental Sciences, University of Colorado: Boulder, CO, 1984; p 248. (9) Dong, T. Y.;Hendrickson, D. N.; Felthouse, T. R.; Shieh, H. S.J. Am. Chem. SOC.1984, 106, 5373-5375. (10) Bencini, A.; Benelli, C.; Gattwchi, D.; Zanchini, C. J. Am. Chem. Soc. 1984, 106, 5813-5818. (1 1) Beck, W. Inorg. Chim. Acta 1985, 99, L33-L34. (12) Porter, L. C.; Doedens, R. J. Inorg. Chem. 1985, 24, 1006-1010. (13) Benelli, C.; Gattwchi, D.; Carnegie, D. W.; Carlin, R. L. J . Am. Chem. SOC.1985, 107, 2560-2561. (14) McCall, D. T.; Doedens, R. J., unpublished work.

actions, characterized by room-temperature diamagnetism, have been found in a series of fivecoordinate copper(I1) complexes with short metal-nitroxyl distance^!**^'*^^^ Very weak couplings, with no measurable effect on magnetic behavior at ordinary temperatures, have also been ~ b s e r v e d . ~ * ~The J ~ intermediate J~J~ coupling regime, in which room-temperature magnetic susceptibility is determined by a distribution over more than one spin state, has had the fewest well-characterized examples. A series of cobalt(I1) halide complexes of di-tert-butyl nitroxide appear to fall into this category, but quantitative interpretation of their magnetic properties has been hampered by instability.'," We now wish to report the preparation, crystal structures, and magnetic properties of a pair of bis(nitroxy1) adducts of Mn(II), which, in addition to providing the best characterized cases of metal-nitroxyl coupling in the intermediate range, are the first reported systems in which two nitroxyl radicals are coordinated to a single metal ion. Experimental Section Synthesis. Mn(F,acac), was prepared from the metal acetate by a (tempo) standard p r d u r e . 1 6 The 2,2,6,6-tetramethylpiperidinyl-l-oxy ligand was used as purchased from Aldrich Chemicals. The ligand 2,2,5,5-tetramethylpyrrolidinyl1-oxy (proxyl) was synthesized from the corresponding crude hydr~xylamine'~-'~ by air oxidation in aqueous (15) Porter, L. C. Ph.D. Thesis, University of California, Irvine, CA, 1984. (16) Cotton, F. A.; Holm, R. H. J. Am. Chem. SOC.1960,82,2979-2983. (17) Shechter, H.; Ley, D. E.;Zeldin, L. J. Am. Chem. SOC.1952, 74, 3664-3668. Todd, (18) Bonnett, R.; Brown, R. F. C.; Clark, V. M.; Sutherland, I. 0.; A. J. Chem. SOC.1959, 2094-2102.

0020-1669/86/1325-2595$01.50/0 0 1986 American Chemical Society